Preparation of Amylose Derivatives Selectively Modified at C-6. 6

Macromolecules 1994,27, 2917-2922

2917

Preparation of Amylose Derivatives Selectively Modified at (2-6. 6-Amino-6-deoxy amylose A. L. Cimecioglu,tJ D. H. Ball;J D. L. Kaplan,t and S. H. Huang*

Biotechnology Diuision, U.S. Army Natick RD&E Center, Natick, Massachusetts 01760,and Biodegradable Polymer Research Consortium, Institute of Materials Science, U-136, University of Connecticut, Storrs, Connecticut 06269 Receiued December 29, 1993; Revised Manuscript Received March 1,1994. ABSTRACT The synthesisof variousamylosederivativeeselectively modified at C-6 leading to the preparation of 6-amino-&deoxyamylose(4) was carried out under homogeneous conditions. Amylose was initially halogenated directly at the primary carbon either by using methanesulfonyl chloride in dimethylformamidel lithium chloride as solvent, giving 6-chloro-6-deoxyamylose(1) or with triphenylphosphine and Nbromoauccinimide in dimethylformamide/lithiumbromide, giving 6-bromo-6-deoxyamyloe(2). Several of these derivatives with different degrees of substitution were prepared. C-6 chlorinated amyloses were then converted to the corresponding 6-azido-6-deoxyamyloaeanalogs (3) by chloride displacement with azide ion in dipolar aproticmedia. Triphenylphosphinefacilitatedreductionofthese intermediatesin dimethylsulfoxide gave 6-amino-6-deoxyamylosea with the same degreesof substitutionasthe C-6chlorinatedprecursors. Producta were characterized m terms of the site and the extent of modifications using ‘3CNMR, FTIR, HPLC, and elemental analyses. Introduction Over the last decade, there has been a growing interest in selective chemical modification of polysaccharides under homogeneous conditions. Such procedures offer several advantagesover classical derivatization methods that are often conducted in a heterogeneous manner, resulting in nonuniform substitutions. Selective approaches, on the other hand, allow control of the degree and distribution of the intended modifications. This facilitates systematic studies of effects of the type and extent of modification on structure/property relationships and on the biological properties of the products. A large volume of literature now available in this field has been compiled in several recent reviews.13 Site-selectivemodifications often exploit the reactivity differences between the functional groups on the sugar residues of polysaccharides. This is relatively straightforward when large differencesexist, for example, carboxyl and amine functionalities in the presence of hydroxyl groups. When, however, all functionalities are the same, such as hydroxyls, as is often the case, more elaborate procedures are needed. These usually require several stages involving sequential protection and deprotection techniques. This has been one major disadvantage associated with selectivemethods,and it has hampered their more wide scale utilization. The more abundant and technologically important polysaccharides, such as cellulose and starch,have nevertheless received considerable attention. The basis of the regiospecific derivatization in these polysaccharides lies in the higher reactivity of the C-6 primary hydroxyl group. Steric accessibility of this site either lends itself to direct derivatization or facilitates its protection and deprotection, thus enabling further chemistry at the secondary positions. The synthesis of several selectively modified cellulose and amylose derivatives has been summarized by H ~ r t o n .Among ~ these, regiospecific halogenation is one of the most effective routes for activation of the primary site, facilitatingfurther transformations at this position. Earlier work has often

* Author to whom correapondenceshould be addreesed.

+ U.S.Army Natick RD&E Center. t University of Connecticut. e Abstract published in Advance ACS

Abstracts, April 16,1994.

0024-929719412227-2917$04.50/0

employed the displacement of a good leaving group, such as tosyl, by the halide ion.6 Recently, several methods, developed in carbohydrate chemistry for the direct replacement of the primary hydroxyl by a halide, have also found applications in polysaccharide modification. One of the most facile of these was described by Evans et a2.6 and involved the use of methanesulfonylchloride in Nfl-dimethylformamide (DMF). This procedure was later effectively extendedto cellulose and am~lose.~ Initial use of this reaction with amylose reported direct displacements of the primary hydroxyl by chlorine up to a DS of 0.9, in spite of the heterogeneous conditions’ (DSis defined here based on substitution at C-6 per anhydroglucoseunit). Nearly complete bromination of amylose at C-6 was also reported by Takeo et alS8using the analogous methanesulfonyl bromide in DMF. These C-6 halogeno amyloses were then employed to prepare other derivatives such as 6-deoxy, 6-azido-6-deoxy-, and 6-aldehydoamylose. Recently, similarly modified pullulans were prepared during an investigation of the effects of substitutionat C-6 on the enzymatic degradation of this polysa~charide.~ We have been studying regiospecific C-6 modifications of amylose with particular emphasis on the preparation of its 6-amino-6-deoxy derivative. 6-Azido-6-deoxyamylose was the intermediate of choice since it appeared to provide a convenient route to the amino derivative by way of reduction. Even though this route is regularly employed in monosaccharidelo and cyclodextrin chemistry,ll its utilization for polysaccharides has been rare. While the preparation of 2-amino-2-deoxyamylose with varying degrees of substitutionhas been reported,1”14 preparation of the 6-amino analog met with limitedsuccess.15 6-Amino6-deoxyamylose with a high degree of substitution could have very interesting properties and potential uses as a cationic polysaccharide. In this article, we report the preparation of several C-6 modified amylose derivatives at varying DS’s up to near complete substitution (DS = l), including 6-amino-6-deoxyamylose. Experimental Section Materials. Amylose (type I11 from potato), methanesulfonyl chloride (Sigma), triphenylphosphine (PbP),N-bromosuccinimide (NBS),lithium chloride, lithium bromide, sodium azide,

1994 American Chemical Society

2918 Cimecioglu et al.

Macromolecules, Vol. 27, No. 11, 1994

Table 1. 'SC NMR Chemical Shifts. of C-6 Modified Amyloses c - 1 c-2c C-3c C-4 C-5 C-6 polymerb solvent amylose DMSO-ds 100.0 71.8 71.8 78.6 73.1 60.4 1 DMSO-& 99.9 71.6 72.4 79.8 69.9 45.3 ~~~

2 3 4

DMSO-de

99.9 100.0 CD3COODd 97.1

DMSO-ds

71.8 71.5 71.1

72.2 72.5 71.7

81.0 79.6 76.6

~~~

69.6 69.7 66.9

c-3

,, c-2

n

~

35.2 51.1 40.6

Ppm from TMS. See Experimental Section for preparation conditions. c Tentative assignments. d 2% in D2O. and lithium azide (Aldrich) were used without further purification. Reaction solvents NJV-dimethylformamide (DMF), dimethyl sulfoxide (DMSO), and N-methylpyrrolidone (NMP) (Aldrich) were anhydrous grade; other solvents were reagent grade, and all were used as received. Analyses. '3c NMR spectra (50.3 MHz) were recorded with a Varian XL-200 spectrometer from DMSO-de solutions with tetramethylsilane (TMS) as the internal standard and occasionally from D2O and 2 % CHsCOOD in DzO solutions with external TMS standard. IR spectra were obtained from KBr d i m using a NicoletPOSXB, fouriertransform spectrometer. HPLC analyees of the hydrolysateswere performed on a Waters AssociatesModel 6000A liquid chromatograph equipped with a YMC-Pack PA 13-5 (250 X 6 mm i.d.) column and a R 4 O l differential refractometer detector. Acetonitrile/water (75/25,v/v) was the mobile phase at a flow rate of 2 mL/min. 6-Chloro-6-deoxyglucaseand glucose were used as standards. Elemental analyses were carried out by Galbraith Laboratories, Knoxville, TN. Syntheses. 6-Chloro-6-deoxyamylose (1). Amylase (20 g, 0.12 mol as anhydroglucose) and lithium chloride (25 g) (dried overnight at 100 "C in vacuum) were dissolved by stirring in DMF (700 mL) at 75 "C (bath temperature) under nitrogen (-2 h). To this homogeneous solution, maintained at 75 "C, methanesulfonyl chloride (100 mL, 1.3mol) was added dropwise over a 15-min period. A white swollen masa precipitated about halfway through the addition. Stirring was continued at this temperature for 2.5 h during which time the mixture became homogeneous and yellowish-brown. After cooling to room temperature, the solution was slowly poured into 3 L of crushed ice-water with vigorous stirring, and the pH of the mixture was adjusted to 8-9 by slow addition of solid sodium carbonate. The suspensionwasstirred Overnight,and the precipitate was collected by fiitration, washed extensively with water and finally with ethanol, and dried overnight at 60 "C in vacuum: yield 20 g (93% based on 6-chloro-6-deoxyamylose). A small amount of the product (- 25 mg) was hydrolyzed by boiling with 2 N aqueous HCl (25 mL) under reflux. The hydrolysate, analyzed by HPLC, contained 6-chloro-6-deoxyglucose with no siflicant quantities of glucose,indicating nearly complete C-6 substitution, DS = -1 was also confiied by the 13C NMR spectrum (Figure 1) and elemental analysis given in Table 2. Partially C-6 Chlorinated Amyloses ( l a and lb). The above described procedure was repeated twice under slightly differing conditions as follows. A homogeneous solution of amylose (8.1 g, 50 mmol as anhydroglucose) and LiCl (10 g) in DMF (200 mL), obtained initially by heating to 80 "C for 2 h under nitrogen, was treated by dropwise addition of methanesulfonyl chloride (39 mL, 0.5 mol) at room temperature. A slight exotherm developed during the addition step and some precipitation also took place as mentioned above. The resulting mixture was atirred at room temperature for 2 days, and the product was recovered and analyzed as described above: la, yield 5.87 g, DS = 0.37, from HPLC of the hydrolysate. A second reaction, also performed at room temperature under similar conditions using amylose (10.4 g, 64 mmol as anhydroglucose), LiCl (12 g), DMF (300 mL), and methanesulfonyl chloride (50 mL, 0.64 mol) for 4 days, gave lb: yield 10.9 g, DS = 0.22, from HPLC of the hydrolysate. 6-Bromo-6-deoxyamylose (2). Amylose (5.1g, 31.5 mmol as anhydroglucose) and lithium bromide (6 g) (dried at 100 "C overnight in vacuum) were dissolved in DMF (200 mL) at 70 "C with stirring under a nitrogen atmosphere until homogeneous ( - 2 h). The solution was cooled to room temperature, and a

n

C-3,2

c-5

"

sb sb B do Figure 1. 1sCNMRspectra of amylose,6-chloro-6-deoxyamylose (l),and partially C-6 chlorinated amylose derivatives (la and lb) (in DMSO-&. ,

'

)

,

I

*do

~

'

~

~

~

-

I

~~ ~

~ ' ~

' I1

~

~, ~~,

I

~

, ~ I

~S

I~ . 'I ~I PA(

Table 2. Elemental Analyses of C-6 Modified Amyloses foundbf polymep %C %H %N % Br/Cl 1

la lb 2 2a 2b 3

3a 3b 4 4a

37.59 (39.91) 39.51 40.53 33.17 (32.02) 34.08 37.40 37.47 (38.51) 38.18 39.55 42.27 (44.72) 37.29 40.00

5.34 (5.02) 6.28 6.43 4.36 (4.03) 4.62 5.31 5.16 (4.85) 6.41 6.39 7.23 (6.88) 6.75 6.57

18.41 (19.63) 7.94 4.90 28.61 (35.51) 20.82 17.94 20.47 (22.45) 8.30 4.70 7.35 (8.69) 3.65 2.44

4b See Experimental Section for preparation conditions. * For amylose,calcd % C 44.45, % H 6.22. e Values in parentheses are that of calculated values for the correspondingcompletelyC-6substituted derivatives. 2-fold exceas/anhydroglucoseunit of NBS (11.2 g, 63 mmol) was added. The resulting yellow solution waa then treated by dropwise addition over a 15-min period of a solution of Phsp (16.5 g, 63 mmol, %foldexcess/anhydroglucoseunit) in DMF (30 mL). A gelatinous precipitate formed at this stage, and the mixture progressively darkened in color. After stirring at room temperature for 30 min, the temperature was raised to 70 OC (bath temperature), and stirring was continued for a further 2 h. The brown and homogeneous mixture was then cooled to a

~ ~

~

Macromolecules, Vol. 27, No. 11,1994 c-3

n

Preparation of Amylose Derivatives Modified at C-6 2919

c-2 nus0

n

c-3

c-7

n I

n

,,

Figure 2. I3CNMR spectra of 6-bromo-6-deoxyamylose(2) and partially C-6 brominated amylose derivatives (2a and 2b) (in DMSO-de). room temperature and slowly poured into 2 L of ice-water containing 50 g of Na2COs. After stirring overnight, the light brown precipitate was collected by filtration, rinsed extensively with water and finally with ethanol, and dried in vacuum at 70 O C : yield 6.5 g (92% based on 6-bromo-6-deoxyamylose). Partially C-6 Brominated Amyloses (2a and 2b). The above describedbromination of amylose was repeated twice using different molar ratios of reactants as follows. 2a: amylose (5.6 g 34.5 mmol as anhydroglucose),NBS (6.14 g, 34.5 mmol), Phsp (9.05g,34.5mmol),LiBr (6g),andDMF(220mLoftotalvolume); yield 6.2 g. 2b: amylose (5.8 g, 35.8 mmol as anhydroglucose), NBS (3.18 g, 17.9 mmol), Phsp (4.69 g, 17.9 mmol), LiBr (6 g), and DMF (220 mL of total volume); yield 3.2 g. 6-Azido-6-deoxyamylose(3). To asolution of 1(4 g, 22 m o l ) in N-methylpyrrolidone (150 mL) was added lithium azide (10 g, 0.2 mol), and the mixture (LiN3 was partially soluble) was stirred at 70 "C for 65 h under nitrogen. The cooled mixture was slowlypoured into ice-water, and the tan colored precipitate was collected by filtration, washed with water and ethanol, and dried in vacuum at 70 O C overnight: yield 3.9 g (95% based on 6-azido6-deoxyamyloae). The lac NMR spectrum (Figure 3) and elemental analysis (Table 2) indicated that completesubstitution by azide ion had been achieved. Partial C-6 Azidodeoxyamyloses(3a and 3b). In a slightly different approach from the above procedure, DMSO solutions (25 and 200 mL) of la and l b (2 and 5 g, respectively) were treated with sodium azide (5g) at 70-75 "C for 48 h. The resulting solutions were diluted with water and dialyzed for 48 h, and the resulting suspensions were lyophilized. 3a: yield 1.5 g. DS of the product was the same as the parent la (0.37) confirmed by the absence of a signal for CH&l in the lSC NMR spectrum (Figure 3) and by the elemental analysis (Table2). 3b: yield 4.6 g; DS = 0.22, as in precursor lb, from 1% NMR (Figure 3) and the elemental analysis (Table 2). 6-Amino-6-deoxyamylose(4). A DMSO solution (120 mL) of 3 (3.6 g, 19.2 mmol) was treated with a 2-fold molar excess of P h a (10g, 38.5 mmol) at room temperaturefor 24 h with stirring. Steady N) evolution subsided after about 2 h. Water (1mL) was then added, and the slightly turbid solution was stirred for a further 24 h. The product was recovered by precipitation into ethanol, filtered off, Soxhlet extraded with ethanol for 3 days, and dried in vacuum at 60 O C : yield 3.1 g (100% based on 6-amino-

I

id0

I

I

I

,

,

,

,

,

,

i o

~

,

1

,

,

,

1

,

I , I I , I I I I

A

I

I

I

I

I

I

I

I

I

I

4b

I

I

I

I

b

I

1

I

I '

2PPu

Figure 3. 13C NMR spectra of 6-azido-6-deoxyamylose(3) and partially C-6 azidated amylose derivatives (3a and 3b) (in DMSOde). 6-deoxyamylose). The 13C NMR spectrum (Figure 4) and elemental analyses (Table 2) confirmed complete reduction of the azido group. Partial C-6 Aminodeoxyamyloses (4a and 4b). The above described reduction with Phsp was repeated for 3a and 3b under similar conditions, except that the initial Phsp treatment was restricted to 3 h. The products were recovered by dialysis (48 h) and freeze dried. Each was finally purified by Soxhlet extraction with ethanol for 3 days. 4a:3a (0.6 g) in DMSO (25 mL) was treated with Phsp (0.8 9): yield 0.53 g, DS = 0.37 as in NMR (Figure 4) and elemental precursors la and 3a from analysis (Table2). 4b3b (4.4 g) in DMSO (150 mL) was treated with Phsp (4 9): yield 4.1 g, DS = 0.22 as in precursors l b and 3b confirmed by NMR (Figure 4) and elemental analysis (Table 2).

Results and Discussion During the synthesis of 6-amino-6-deoxyamylose (4), several C-6 modified intermediates were prepared. The route taken is illustrated in Scheme 1. The first stage involved halogenation at the primary carbon to give 6-chloro- and 6-bromodeoxyamyloses (2,2a and 2b) with different degrees of substitution. The 6-chloro derivatives (1, la, and lb) underwent nucleophilic substitution to give the 6-azido analogs (3, 3a, and 3b, respectively), and, finally, the reduction of this set yielded the 6-amino-6deoxyamyloses (4, 4a, and 4b respectively) which had essentially the same DS as the parent chlorinated derivatives. Chlorination at the primary carbon of amylose was

selected as the initial activation step since it has been shown to proceed with a high degree of selectivity. This is unlike other routes such as toaylation and meaylation which at high degrees of substitution have been found to suffer in selectivity and also involve reactions in secondary position^.^ While the literature on direct C-6 chlorination of amylose is limited, there are several examples for cellulose.lem The reagents commonly employed are thionyl chloride, methanesulfonyl chloride, chlorodi-

2920 Cimecioglu et al.

Macromolecules, Vol. 27, No. 11, 1994

Scheme 1. The Synthesis of C-6 Modified Amylose Derivatives CHjSOzCl DMFLiCl 75'CIZSh

-

Ph3PrnS DMFiLiBr 700CI2h

LiN3MMP M NaNGMSO 70-75W4845h

I

2

o\

o\

3

1. Ph3PiDMSO

RTR4h 2. H20 RTt24h

r CH2NH

to remove even after extensive drying. Additional confirmation of the selective C-6 chlorination was found in the 13CNMR spectra shown in Figure 1. Upon replacement of the primary hydroxyl with chlorine the C-6 resonance shifts upfield to 6 = 45.3 ppm from 6 = 60.4 ppm (Table 1). While the partial, but increasing chlorination level was indicated in Figure 1 for l b and la, no significant resonance corresponding to the primary CH2OH was visible in the spectrum of 1. This, together with HPLC analysis, confirmed the near complete C-6 substituted nature of 1. The very high degree of modification was also reflected in the solubility behavior of 1, since it was soluble in most dipolar aprotic solventssuchas DMSO, DMF, DMA, and NMP as well as pyridine. Partially chlorinated derivatives l a and lb, on the other hand, showed more complex solubility characteristics. The former was partially soluble in water, and the latter was not, which is believed to account for the lower yield obtained from the reaction giving la. Similar solubility changes for the partially C-6 chlorinated amyloses were also reported by Horton et aL7 Different DS's exhibited by the products l a and lb, even though they were prepared under essentially the same conditions, i.e., at room temperature, can probably be explained by the mechanism of this type of reaction which is well understood.20 Following the reaction of methanesulfonylchloride with DMF to give 0-(methanesulfony1)-NJV-dimethylformiminium salt,CH3SO&H=NMe2+Cl-, two subsequentroutes are possible. The first yields the amylose, NJV-formimidium chloride, Amy-O-CH=NMe2+Cl-, by reaction with an amylose hydroxyl, or the second which yields chloro NJV-dimethylformiminium chloride, ClCH=NMe2+Cl-, by chloride ion displacement. An amylose hydroxyl can then displace chlorine in the latter case; again yielding Amy-O-CH=NMe2+Cl-. This intermediate ionic product is believed to be the precipitate that forms during methanesulfonylchloride addition in our reactions. Two further routes are then possible: hydrolysis with water to give the formyl ester or chloride ion displacement at elevated temperature to give chlorodeoxyamylose. Thus the different DS's observed with la and l b which were obtained after 2-4 days at room temperature are most probably attributable to the fluctuations in temperature during the long course of these treatments and, perhaps, to a lesser extent, to the partially heterogeneous nature of the reactions. For various reasons it was also desirable to have a better leaving group than chlorine at the primary position of amylose. 6-Bromo-6-deoxyamylosewas the obvious choice, provided that it could be prepared as selectively as the chloro derivative. The only mention of 2 in the literature was made by Takeo et aL8 who used methanesulfonyl bromide in DMF to successfully brominate amylose to high DS (0.92). This reagent was not, however, readily accessible, which prompted us to seek an alternative procedure. Of the several methods surveyed in carbohydrate chemistry for the direct replacement of primary hydroxyl groups, the use of PhsP in the presence of N-halosuccinimides in DMF has been shown to be one of the most facile in terms of the selectivity.21 The reaction proceeds to high yields under essentially neutral conditions.21 A similar approach was used for 6-halogeno-6deoxycellulom preparations with N-halaeuocinimideswhere secondary hydroxyls were protected during the reaction.22 Apart from this report, no systematic use of these reagents appears to have been made in polysaccharide modifications. Boger et a1.* have, nevertheless, successfully employed PhP with carbon tetrabromide in the presence of LiN3to prepare 6-azido-6-deoxy-a-and &cyclodextrins

Hq 4

O,

methylformimidium chloride, and p-toluenesulfonyl chloride. DMF has been the solvent used almost exclusively. Most of these reactions, however, were carried out under heterogeneous conditions that undoubtedly resulted in nonrandom modification,except at the highest degree of substitution (which is often unattainable due to the heterogeneous conditions). Not until solvent systems capable of dissolving cellulose such as N,N-dimethylacetamide (DMA)/LiClor DMF/chloralhas it been possible to effect these halogenations under homogeneous conditions.lgm To our knowledge, however, no similar attempts were made with amylose. In this respect, we have decided to perform the C-6 modifications of amylose in a DMF/LiCl solvent system. When up to 5 wt% of amylose was treated with DMF containing 4-5 wt% LiCl at 70-80 OC,homogeneous solutions were obtained within 2 h, which upon cooling to room temperature, remained clear. Initially, these solutions were treated by dropwise addition during 10-20 min of a 10-fold molar excess/ anhydroglucose residue of methanesulfonyl chloride. A slight exotherm developed and, depending on the rate of addition of the reagent, a highly swollen gelatinous mass precipitated, usually halfway through the addition step. The mixtures were then stirred at room temperature for several days and gave partially substituted products la and lb. A third reaction, carried out at 75 OC throughout, was observed to rehomogenize within 1.5-2 h after the addition of the reagent and was discontinued after 2.5 h, giving the product 1. Upon hydrolysis of these products only glucose and 6-chloro-6-deoxyglucose was obtained. Quantitative HPLC analysisof the hydrolysatesindicated DS's of 1, 0.37, and 0.22 for 1, la, and lb, respectively. These values were also tentatively confirmed by elemental analyses given in Table 2. The slightly lower elemental compositions observed were attributed to significant moisture content of the samples which is often difficult

Macromolecules, Vol. 27, No. 11,1994 c-2

Figure 4. 13C NMR spectra of 6-amino-6-deoxyamylose(4) and partially C-6 aminated amylose derivatives (4a and 4b) (4 and 4b in 2% CDgCOOD in DzO, 4a in DzO).

with high degrees of selectivity. As in the case of LiC1, the use of LiBr in DMF also facilitated the dissolution of amylose at reasonable concentrations (C5 wt%). These solutions were then treatedwith different molar equivalents of P b P and NBS accordingto the procedure described in the Experimental Section. Products 2, 2a, and 2b were obtained by treatment with 2-fold, equivalent and V2-fold molar proportion of the reagents per anhydroglucose unit, respectively. l3C NMR spectra of the products are given in Figure 2, and the elemental analyses are listed in Table 2. It was apparent that selective bromination of the primary carbon occurred as the CHzOH resonances at 6 = 60.4 ppm were replaced by a new peak at 6 = 35.2 ppm (Table 1).As inferred from its 13CNMR spectrum, nearly completeC-6 bromination of amylose (product 2) resulted by utilizing a 2-fold molar excess of the reagents. Elemental analyses of this particular sample, however, had significantlylower bromine content than expected. This is believed to be a consequence of a small amount of gelatinous residue that was often observed during the dissolution of the sample in DMSO. Therefore, it is conceivablethat a side reaction might have come into effect at the higher substitution attempt, leading to, perhaps, some cross-linking and hence resulting in the observed lower bromine content of the sample. This was not the case for 2a and 2b. However, in light of the slightly lower than expected elemental analyses obtained from these polymers, due to the reasons mentioned earlier, at this stage only a rough estimation of the DS of these samples was possible. Thus, the C-6 bromination levels in 2a and 2b are estimated to be around 60 and 40%, respectively. It was also noted that the 13Cspectrum of 2a (Figure 2) contained an additional shoulder in the anomeric carbon region at ca. 6 = 97 ppm. This resonance is believed to be due to a small amount of 3,g-anhydro ring formation, which most likely occurred during the workup of the polymer. 3S-Anhydro ring formation is a common oc-

Preparation of Amylose Derivatives Modified at C-6 2921 currence in carbohydrates containing a leaving group in the primary position (or the secondary positions for 2,3and other anhydro derivatives). Since slightly basic conditions were employed during the workup (see Experimental Section), the presence of bromine on C-6 presumably increased the likelihood of internal cyclization even under heterogeneous conditions. However, as was the case for 2b, this could be avoided by careful control of the pH of the supernatant during precipitation. Otherwise the reactions proceeded smoothly, giving somewhat more discolored products at higher bromination levels. An intermediate precipitation and redissolution during the addition stage also occurred for 2. This was much less pronounced for 2a and was absent in case of 2b. The precipitate is believed to be the alkoxyphosphonium bromide salt, Amy-O-PPb+Br,from which triphenylphosphine oxide is displaced by the bromide upon heating, as described elsewhere in the mechanistic analysis of similar reactions.24 It thus appears that the PhsP/NBS/DMF/ LiBr system is equally effective for the selective halogenation at the primary carbon of amylose and should also be applicable to other polysaccharides. 6-Chloro-&deoxyamylose derivatives, 1, la, and l b discussed earlier, were converted to the corresponding 6-azido analogs, 3, 3a, and 3b, using well established reactions. Two routes employed at this stage involved either the use of LiN3 in NMP or NaN3 in DMSO. While the former reaction was heterogeneous due to the partial solubility of LiN3 in NMP, the latter was homogeneous throughout. Products, analyzed by l3C NMR spectroscopy (Figure 3, CH2N3; 6 = 51.1 ppm, Table l),and elemental analyses (Table 2) were found to be completelysubstituted by the azide with no significant amounts of CHzCl groups remaining. The IR spectrum of 3 (Figure5 ) also confirmed the conversion,as the C-Cl stretch at 662 cm-l was replaced by the C-N3 stretching vibration at 2107 cm-l. 6-Azido6-deoxyamylosederivatives were also accessible through C-6 brominated precursors (2,2a, and 2b) via treatment with NaN3 in DMSO, as illustrated in Scheme 1. In this case, reactions proceeded under milder conditions, at 50 OC, and were complete within 6 h. However, due to the lesser degree of certainty concerning the extent of modification of the C-6 brominated intermediates, this route was not exploited any further at this stage. In the final stage of the sequencesillustrated in Scheme 1,6-azido derivatives were reduced to the corresponding 6-amino-6-deoxyamyloseanalogs 4, 4a, and 4b. Traditionally, the amino derivatives of polysaccharides are usually prepared by catalytic reduction of oxime or azide precursors. Difficulties in these reactions are often encountered due to limited solubilities of the polysaccharides which can result in incomplete conversion or in low overall yields. Teshirogi et a1.15 have reported the synthesis of 6-amino-6-deoxystarchby sequential tosylation, azide displacement, acetylation, and LiAlH4 reduction steps. Even though a DS of 0.92 was obtained, the overall yield was only 145%. For our reduction, we elected to attempt a relatively little used procedure, at least in carbohydrate chemistry, which involves the reaction of a primary azide with Ph3P followed by the hydrolysis of the iminophospharene formed, according to the following scheme. RN,

+ Ph3P

-N2

RN=PPh,

Ha0

RNH,

+ Ph3P0

Also known as the Staudinger this method has recently been demonstrated to promote quantitative reduction of primary azides to the correspondingprimary

2922 Cimecioglu et al.

Macromolecules, Vol. 27, No. 11, 1994 in water, 4b was not, but it dissolved in dilute acetic acid and DMSO. The 13CNMR spectra (Figure 4), showed a new resonance at 6 = 40.6 ppm (Table l),and in all cases the peaks corresponding to the primary azido carbon (6 = 51.1 ppm, Figure 3) were absent. Elemental analyses (Table 2) also indicated that reduction of the azides had taken place. The DS's of the products were, therefore, essentially the same as those of the parent 6-azido and 6-chloro derivatives, i.e., -1,0.37, and 0.22 for 4,4a, and 4b, respectively. Furthermore, as also described in the literature,26two stage treatment of the azide could be avoided in our system by adding Ph3P and water simultaneously, each in slight excess, to the DMSO solution of the azide. Ph3P facilitated reduction of primary azides, demonstrated here for the preparation of 6-amino-6-deoxymylose, was found to be extremely facile and proceeded to completion under neutral conditions in DMSO a t room temperature. The great utility of this method should no doubt find further application in the amination of other polysaccharides.

References and Notes (1) Yalpani, M. Polysaccharides; Elsevier: Amsterdam, 1988. (2) Yalpani, M. Tetrahedron 1985,41, 2957. (3) Beck, R. H. F.; Fitton, M. G.; Kricheldorf, H. R. Handbook of

~-4000

V

__

_.

30W

2000

-

1000

YAVENIJNEER (crn-')

Figure 5. FTIR spectra of amylose, 6-chloro-6-deoxyamylose (1),6-bromo-6-deoxyamylose (2),6-azido-6-deoxyamyloe(3,and 6-amino-6-deoxyamylose (4).

amines in polar solvents such as THF and at room temperature.26 Earlier Boger et aLZ3utilized this reaction to effect the reduction of hexakis(6-azido-6-deoxy)-olcyclodextrin to its 6-amino analog in quantitative yield in dioxane. Our initial attempts at the reduction of 3 in THF and NMP was only partially successful primarily due to the heterogeneous reaction in the former solvent and to a somewhat reluctant transformation even after prolonged treatment in the latter case. When these two solvents were used together (1:l v/v), better results were obtained, but the IR spectrum of the product still indicated the presence of azide. In all cases,further characterization of the products was hampered by their insolubility, once recovered,in any solvent, even though they did swell highly in dilute acetic acid solutions. However, when a solution of 3 (ca. 3 w t % ) in DMSO was treated with a 2 molar excess of Ph3P, as described in the Experimental Section, an immediate evolution of gas bubbles occurred which subsided after 2-3 h. The product (4) recovered, after hydrolysis, was insoluble in all solvents, swellingin DMSO and water. It, nevertheless, dissolved in 2% acetic acid up to 1 wt% . Concentrations above this level gave either extremely viscous solutions or gelled. The IR spectrum (Figure 5) of 4 no longer contained the azide stretching band at 2107 cm-'. Products 3a and 3b were also treated in a similar manner to give 4a and 4b. While 4a was soluble

Polymer Synthesis; Kricheldorf, H. R., Ed.; Marcel Dekker: New York, 1992. (4) Horton, D. New Developments in Industrial Polysaccharides; Crescenzi,V., Dea, I. C. M.; Stivala, S. S., Eds.; Gordon Breach New York, 1985. (5) (a) Tipson, R. S. Ado. Carbohydr. Chem. 1953,8,107. (b) Ball, D. H.; Parrish, F. W. Adv. Carbohydr. Chem. Biochem. 1969, 24,139.

(6) Evans, M. E.; Long, L., Jr.; Parrish, F. W. J. Org. Chem. 1968, 33, 1074. (7) Horton, D.; Luetzow, A. E.; Theander, 0.Carbohydr. Res. 1973, 27, 268. (8) Takeo, K.; Sumimoto,T.;Kuge, T.Starke 1974,26,111. (9) Ball, D. H.; Wiley, B. J.; Reese, E. T. Can. J. Microbiol. 1992, 38, 324. (10) Horton, D. The Amino Sugars; Jeanloz, R. W., Ed.; Academic Press: New York, 1969; Vol. IA. (11) Croft, A. P.; Bartach, R. A. Tetrahedron 1983,39, 1417. (12) Wolfrom, M. L.; Taha, M. I.; Horton, D. J. Org. Chem. 1963, 28,3553. (13) Wolfrom, M. L.; Kato, H.; Taha,M. I.; Sato, A.; Yuen, G. U.; Kinoshita, T.;Soltes, E. J. J. Org. Chem. 1967,32,3086. (14) Wolfrom. W. L.: Wane, P. Y. Carbohvdr. Res. 1970.12. 109. (15) Teahirod, T.;Y k m & o , H.; Sakamoh, M.; Tonami; H. ken? Gakkakhi 1979.35, T479. (16) Vigo, T. L.; Daig1e;D. J.; Welch, C. M. J. Polym. Sci. Polym. Lett. Ed. 1972, 10, 397. (17) Horton, D.; Luetzow, A. E.; Theander, 0. Carbohydr.Res. 1973, 26, 1. (18) Ishii, T.; Ishizu, A.; Nakano, J. Carbohydr.Res. 1977,59,155. (19) McCormick, C. L.; Callais, P. Polymer 1987,28, 2317. (20) McCormick, C. L.; Dawsey, T. R.; Newman, J. K. Carbohydr. Res. 1990,208, 183. (21) Hanessian, S.;Ponpipom, M. M.; Lavalle, P. Carbohydr. Res. 1972,24,45. (22) Gal'braikh,L. S.; Rogovin,Z. A. Cellulose Chem. Technol. 1968, 2, 375. (23) Boger, J.; Corcoran,R. J.; Lehn, J.-M. Helu. Chim. Acta 1978, 61, 2190. (24) Hodosi, G.; Podanyi, B.; Kuszmann, J. Carbohydr. Res. 1992, 230, 327. (25) Scriven, E. F. W.; Tumbull, K. Chem. Rev. 1988,88,297. (26) Knouzi, N.; Vaultier, M.; Carrie, R. Bull. SOC.Chim. Fr. 1986, 815.